EP2348410B1 - Virtual-CPU based frequency and voltage scaling - Google Patents

Virtual-CPU based frequency and voltage scaling Download PDF

Info

Publication number
EP2348410B1
EP2348410B1 EP10252026.9A EP10252026A EP2348410B1 EP 2348410 B1 EP2348410 B1 EP 2348410B1 EP 10252026 A EP10252026 A EP 10252026A EP 2348410 B1 EP2348410 B1 EP 2348410B1
Authority
EP
European Patent Office
Prior art keywords
workload
virtual
scaling algorithm
cpu
scaling
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Not-in-force
Application number
EP10252026.9A
Other languages
German (de)
English (en)
French (fr)
Other versions
EP2348410A2 (en
EP2348410A3 (en
Inventor
Kun TIAN
Ke Yu
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Intel Corp
Original Assignee
Intel Corp
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Intel Corp filed Critical Intel Corp
Publication of EP2348410A2 publication Critical patent/EP2348410A2/en
Publication of EP2348410A3 publication Critical patent/EP2348410A3/en
Application granted granted Critical
Publication of EP2348410B1 publication Critical patent/EP2348410B1/en
Not-in-force legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F9/00—Arrangements for program control, e.g. control units
    • G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
    • G06F9/46—Multiprogramming arrangements
    • G06F9/50—Allocation of resources, e.g. of the central processing unit [CPU]
    • G06F9/5061—Partitioning or combining of resources
    • G06F9/5077—Logical partitioning of resources; Management or configuration of virtualized resources
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F1/00—Details not covered by groups G06F3/00 - G06F13/00 and G06F21/00
    • G06F1/26—Power supply means, e.g. regulation thereof
    • G06F1/32—Means for saving power
    • G06F1/3203—Power management, i.e. event-based initiation of a power-saving mode
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F1/00—Details not covered by groups G06F3/00 - G06F13/00 and G06F21/00
    • G06F1/26—Power supply means, e.g. regulation thereof
    • G06F1/32—Means for saving power
    • G06F1/3203—Power management, i.e. event-based initiation of a power-saving mode
    • G06F1/3234—Power saving characterised by the action undertaken
    • G06F1/324—Power saving characterised by the action undertaken by lowering clock frequency
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F1/00—Details not covered by groups G06F3/00 - G06F13/00 and G06F21/00
    • G06F1/26—Power supply means, e.g. regulation thereof
    • G06F1/32—Means for saving power
    • G06F1/3203—Power management, i.e. event-based initiation of a power-saving mode
    • G06F1/3234—Power saving characterised by the action undertaken
    • G06F1/3296—Power saving characterised by the action undertaken by lowering the supply or operating voltage
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
    • Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management

Definitions

  • Virtualization generally is implemented using a software entity referred to as a Virtual Machine Monitor (VMM).
  • VMM presents to each OS a virtual machine (VM) having virtual resources, including one or more virtual processors, virtual memory and virtual input/output (I/O) resources, that the OS may completely and directly control.
  • the VMM maintains the system environment for implementing virtualization policies, such as sharing and/or allocating physical resources among the VMs.
  • Each OS and other software that runs on a VM is referred to as a guest or guest software, while a host or host software is software, such as the VMM, that runs outside of the virtualization environment.
  • VMM virtual central processing unit
  • a guest executing on a system may see a virtual central processing unit (VCPU) which appears to the guest as a physical CPU having the same, or a subset of the same, features as the actual physical CPU (PCPU).
  • VCPU virtual central processing unit
  • virtualization technologies can consolidate many diverse workloads into a single physical machine, virtualization makes efficient use of physical resources and, thus, in some respects, may be viewed as a type of environmentally friendly (or "green”) technology. This has provided momentum to embrace virtualization technologies, particularly since energy consumption is of increasing concern. However, simply consolidating OSs on a single platform may still not be enough to satisfy demands for efficiency. Towards that end, virtualization technologies are increasingly being required to further exploit fine-grained power management (or power-saving) capabilities of the type that were first introduced on client devices and now are popular even on high end servers. However, virtualization of the CPUs and the continuous sharing and re-allocating of physical resources adds a layer of complexity, making it difficult to translate traditional power management techniques implemented by an OS on a physical machine to a virtual environment.
  • a virtual machine monitor acquires information about loads on virtual CPUs (VCPUs), and the VMM assigns the true CPU resources based on the information about the loads on the VCPUs in such a manner that more true CPU resources are assigned to a VCPU with a higher load.
  • US2006/0047987 discloses techniques for reducing power consumption within computing devices, such as wireless communication devices.
  • a device dynamically adjusts the CPU clock frequency based on CPU load in order to reduce power consumption.
  • the device monitors the load of the CPU using a number of sample interrupts.
  • the device determines whether to adjust the clock frequency based on the monitored load of the CPU. In general, the device increases the clock frequency when the load of the CPU is high and decreases the clock frequency when the load of the CPU is low.
  • power-saving capabilities may be efficiently integrated with virtualization technology.
  • Current virtual systems either fail to provide any type of scaling capabilities (and thus do little in terms of power-saving) or simply replicate the same power-saving algorithms that are implemented by the OS in a physical machine.
  • scaling algorithms that are intended for use in physical environments do not translate well to virtual environments.
  • One example of a general scaling technique implemented in a physical machine is referred to as dynamic voltage and frequency scaling (or DVFS).
  • Dissipated power in a chipset e.g., a processor or CPU
  • DVFS dynamic voltage and frequency scaling
  • a DVFS technique power consumption is efficiently reduced by changing the voltage and frequency of the signals applied to the CPU based upon changing circumstances.
  • scaling algorithms intelligently determine when it is appropriate to scale voltage and frequency in a manner that will have minimal impact on performance.
  • One way to do this is to base the scaling decision on CPU utilization.
  • frequency and voltage scaling techniques are coupled to each virtual CPU rather than to the physical CPU.
  • each virtual CPU is provisioned with its own workload profiling channel and/or its own scaling algorithm to tailor and couple the scaling to that particular VCPU.
  • some embodiments profile workload characteristics based on continuous monitoring of virtualization events to dynamically and more intelligently react to the diverse workloads running on the various VCPUs.
  • multiple different scaling algorithms may be available for selection as a function of the actual characteristics of the currently executing workload. In this manner, frequency and voltage scaling can respond to fast-changing consolidated workload patterns with the result that power savings may be achieved without sacrificing performance.
  • Frequency and voltage scaling algorithms generally are based on historical information regarding the demand for and utilization of the PCPU. For instance, some OS implementations periodically sample PCPU utilization, e.g., in 20 millisecond (ms) intervals. Based on that past pattern of PCPU utilization, an appropriate clock frequency (or scaling factor) may be determined for the CPU in the next 20ms interval. For instance, if CPU utilization is high, then a high frequency will be selected. Likewise, if utilization is low, then a low frequency may be selected. This technique generally is adequate in physical systems because the OS system schedulers typically run in short intervals (e.g., 1ms), and thus workload patterns viewed in windows of 20ms generally are sufficient to provide an accurate indication of an average workload pattern on a single PCPU.
  • short intervals e.g. 1ms
  • VMM schedulers behave differently because of the need to time-share a PCPU for VCPUs running on that PCPU. Rather than running in short intervals, a VMM scheduler relaxes its scheduling of tasks to a larger interval to avoid the high context switch overhead that generally accompanies each newly introduced virtualization layer. For instance, a VMM typically may have a scheduling interval on the order of 30ms. Since this time quantum (e.g., 30ms) is relatively close in size to the frequency scaling interval (e.g., 20ms) of PCPU scaling algorithms, there is a high likelihood that a pattern of the previous scheduled-out VCPU will be used to determine the frequency scaling for the next scheduled-in VCPU.
  • time quantum e.g. 30ms
  • the frequency scaling interval e.g. 20ms
  • the frequency selected for the scheduled-in VCPU may be poorly suited for the work currently being performed.
  • This problem is one type of problem that is referred to as "false scaling.”
  • One way to avoid the false scaling problem in this instance would be to simply enlarge the scaling interval (e.g., to 200ms or more).
  • a large scaling interval in turn may result in a scaling algorithm that cannot timely respond to fast-changing characteristics of ongoing workloads and may also cause fairness issues with sharing resources.
  • embodiments of the invention avoid the false-scaling problem by provisioning each VCPU with its own profiling channel, thus allowing the characteristics of each VCPU to be individually monitored and analyzed.
  • scaling decisions can be based on the characteristics of the workload that is actually executing on the VCPU.
  • scaling decisions typically may be a function of CPU utilization.
  • CPU utilization is not always an accurate indicator for the optimal frequency for the particular workload, which can also result in a false-scaling situation. For instance, a memory intensive workload may saturate the CPU, but may also have many stall cycles. Thus, although CPU utilization would indicate that a high frequency should be selected, a low frequency actually could be selected without detrimentally affecting performance. Accordingly, some embodiments of the invention address this type of false-scaling problem by adapting the workload characteristic profiling channel to recognize workload patterns that are not based on CPU utilization.
  • VCPU-based scaling also may offer further advantages by using different scaling algorithms for each VCPU.
  • multiple scaling algorithms may be available for use by any one VCPU.
  • one scaling algorithm may be suitable for selecting a frequency based on CPU utilization
  • another scaling algorithm may be configured to select a frequency based on performance considerations
  • yet another scaling algorithm may base its decisions on power savings, etc.
  • a particular scaling algorithm may be dynamically selected based on characteristics of the current workload that are provided by the profiling channel.
  • a currently selected scaling algorithm may then be replaced with another scaling algorithm when and if the characteristics of the current workload changes.
  • I/O intensive workloads, memory intensive workloads, and latency sensitive workloads may all benefit from the use of different scaling algorithms that take into consideration the realities of the characteristics of those workloads.
  • VCPU-based scaling In addition to providing improved scaling capabilities that are tailored for each VCPU, VCPU-based scaling also provides opportunities for more accurately determining compliance with service level agreements between the provider of the computing services and its customers. For instance, such agreements may have certain provisions that are based on utilization and speed of the physical processing resources. Thus, the frequency at which each VCPU is running is a factor which may be taken into consideration for the service level agreement requirements. If PCPU-based scaling were used, it might not be possible to accurately ensure compliance with a frequency requirement. Thus, a PCPU-based scaling algorithm may have to be disabled as long as any VCPU that is subject to such a requirement is running on a PCPU. VCPU-based scaling can thus facilitate metering and accounting for utilization of the system's processing resources for each customer's workload in a more accurate manner.
  • Fig. 1 illustrates an exemplary framework 100 for implementing VCPU-based scaling in a virtual environment in accordance with one embodiment of the invention.
  • Fig. 1 shows an embodiment of a virtualization framework 100, it should be understood that embodiments may also be implemented in other frameworks, architectures, systems, platforms or environments depending on the particular application in which the virtualization techniques are employed
  • a virtualization framework 100 includes a bare hardware platform 102 that may correspond to any type of computer system that can execute OS or VMM software.
  • bare hardware platform may be that of a personal computer, mainframe computer, portable computer, handheld computer, set-top box, server, or ay other computing system.
  • the bare hardware platform 102 includes a physical processor (or PCPU) 104, a memory 106, I/O resources 107, and clock generation/scaling circuitry 108 for generating a clock signal having a frequency and a voltage for use by the various components of the platform 102, such as the PCPU 104.
  • PCPU physical processor
  • PCPU 104 may be any type of processor, including a general-purpose microprocessor, such as a multi-core processor, microcontroller, or programmable logic. Although Fig. 1 shows only one PCPU 104, it should be understood that bare hardware platform 102 may include multiple processors, including any number of multi-core processors, each with any number of execution cores, and any number of multithreaded processors, each with any number of threads.
  • the memory 106 shown in Fig. 1 may include dynamic random access memory (DRAM) or other types of memory, such as static random access memory.
  • Bare platform hardware 102 may also include any number of additional devices or connections, including various storage devices for storing data, applications, instructions of software, etc.
  • the system 100 in Fig. 1 includes a VMM 110, and VMs 118 and 120 having VCPUs 112a, 112b, 112n, guests OSs 122 and 124 and guest applications 126, 128, 130 and 134.
  • guest OS 122 and guest applications 126 and 128 are installed in VM 118 for execution by VCPUs 112a and 112b.
  • Guest OS 124 and guest application 130 and 132 are installed in VM 120 for execution by VCPU 112n.
  • system 100 may include any number of VMs, guests, and VCPUs, that the VMs may include other virtual resources (e.g., virtual memory, virtual I/O resources, etc.), and that the scope of the invention is not limited in this regard.
  • VMs may include other virtual resources (e.g., virtual memory, virtual I/O resources, etc.), and that the scope of the invention is not limited in this regard.
  • VMM 110 may be any software, firmware or hardware host installed on or accessible to bare platform hardware 102 to present VMs, i.e., abstractions of bare platform hardware 102, to guests or to otherwise create and manage VMs and implement virtualization policies.
  • a host may be any VMM, hypervisor, OS, or other software, firmware, or hardware capable of controlling bare platform hardware 102.
  • a guest may be any OS, any VMM, including another instance of VMM 110, any hypervisor, or any application or other software.
  • the VMM 110 includes a virtual based frequency and voltage scaling (VBFVS) center 134 and a plurality of VBFVS instances 136a-n.
  • Each VBFVS instance 136a-n is allocated to a VCPU 112a-n.
  • Each VBFVS instance 136a-n also is bound to a profiling channel 138a-n that profiles the characteristics of the VCPU's workload and to a scaling algorithm 140a-n that determines a suitable scaled frequency and/or voltage (or scaling factor) as a function of the profile provided by the profiling channel.
  • the binding to a particular scaling algorithm 140 may be dynamically performed.
  • profiling data may indicate that a scaling algorithm 140b may be better suited for the workload currently running on the VCPU than the currently bound algorithm 140a.
  • the VBFVS instance 136 may dynamically switch from the algorithm 140a to the scaling algorithm 140b.
  • system 100 also includes a VMM scheduler 142 for scheduling each of the VCPUs 112a-n.
  • the scheduler 142 schedules in a new VCPU (e.g., VCPU 112b)
  • the scheduler 142 sends a context switch event to the VBFVS center 134.
  • the VBFVS 134 center accepts the context switch event
  • the VBFVS center 134 selects the VBFVS instance (e.g., instance 136b) to allocate to the scheduled-in VCPU 112b.
  • the profiling channel (e.g., channel 138b) associated with the VBFVS instance 136b is initiated and provides information to the scaling algorithm (e.g., algorithm 140b) bound to the instance 136b to determine an appropriately scaled frequency and/or voltage.
  • the VBFVS center 134 then scales the frequency by communicating an indication of the new frequency to the clock generation circuitry 108 that generates the clock signal for the PCPU 104. In some embodiments, the scaling is performed repeatedly during the entire period during which the VCPU 112b is scheduled in.
  • each VBFVS instance 136 also is associated with an Accounting and Budgeting Module (ABM) 144 to provide for fine-grained accounting of power and energy at the VCPU level.
  • ABSM Accounting and Budgeting Module
  • Fig. 2 shows an exemplary flow diagram for determining frequency and/or voltage scaling that may be implemented in the system 100. It should be understood that the invention is not limited to the particular steps or the particular order of steps shown. Other embodiments may include more or fewer steps and may alter the order in which the steps are performed.
  • the VMM scheduler 142 schedules in a new VCPU (e.g., VCPU 112b).
  • the VMM scheduler 142 sends a context switch event to the VBFVS center 134.
  • the VBFVS center 134 selects the VBFVS instance 136 for the scheduled-in VCPU 112 (block 206).
  • the VBFVS center 134 may also send scaling statistics to the scheduler 142. For instance, the VBFVS center 134 may store which profiling channel and scaling algorithm was last used with the scheduled-out VCPU and may update statistics regarding the workload characteristics and VCPU usage.
  • the VBFVS center 134 may also record the last frequency that had been determined for use by the scheduled-out VCPU.
  • the VBFVS center 134 also may send the statistics to the scheduler 142.
  • the VBFVS center 134 may recover the previously recorded last frequency for the scheduled-in VCPU 112 and thaw the profiling mechanism and scaling algorithm that had last been used with the current scheduled-in VCPU 112.
  • the VBFVS center 134 may use other techniques for determining which frequency, profiling mechanism and scaling algorithm to initially employ with a scheduled-in VCPU, such as using default starting parameters.
  • the selected profiling channel 138 is activated and begins collecting data to profile the characteristics of the currently executing workload on the scheduled-in VCPU 112 (blocks 208 and 210). The profiled information is then provided to the selected scaling algorithm 140.
  • the algorithm 140 determines the appropriate frequency/voltage (or scaling factor) and communicates it to the VBFVS center 134.
  • the VBFVS center 134 then causes the clock generation circuitry 108 to scale the clock frequency/voltage accordingly.
  • the profiling channel 138 continues to collect and send data to the scaling algorithm 140, and the scaling algorithm 140 continues to scale the frequency and/or voltage in response to the data until a new VCPU is scheduled in.
  • the scaling technique shown in Fig. 2 also includes a step of changing or selecting a new scaling algorithm 140 based on the profiled workload (block 214).
  • one type of scaling algorithm that may be available for selection is a CPU-demand-based scaling (CPU-DBS) algorithm.
  • utilization of a physical CPU may be calculated by: T elapsed ⁇ T idle / T elapsed where T elapsed is the elapsed time since utilization was last sampled, and T idle is the CPU idle time during the elapsed phase.
  • this demand-based formula may be modified such that it is closely coupled to the VCPU and, as a result, reduces occurrences of false-scaling. For instance, several new factors may be introduced into the profiling channel 138 that take the virtual environment into consideration. As an example, in virtual machines, the VMM scheduler 142 generally allocates some time quantum (e.g., 30ms) for each scheduled-in VCPU 112. However, there are several situations where a scheduled-in VCPU 112 may not consume all of its granted quantum in one time slice. One such situation occurs when the VCPU 112 itself stops running because it enters an idle state.
  • some time quantum e.g. 30ms
  • the VMM scheduler 142 may place the VCPU 112 in a blocked state until a virtualization event occurs to wake the VCPU 112 up.
  • the time during which the VCPU 112 remains in the blocked state may be referred to as T blocked .
  • the scheduler 142 may be adapted to communicate the granted quantum and the blocked time parameter to the profiling channel 138 for use in profiling the characteristics of the workload.
  • Another situation in which the VCPU 112 may not consume its entire quantum in a single time slice may occur when the VCPU 112 is preempted by other higher priority VCPUs 112.
  • the preempted VCPU 112 is still runnable but is pending in a run queue and actually has some meaningful jobs ongoing.
  • this preempted time may be referred to as T stolen .
  • the scheduler 142 may be adapted to communication the stolen time parameter to the profiling channel 138.
  • the profiling channel 138 may calculate VCPU 112 utilization as follows: T quantum ⁇ T stolen ⁇ T blocked / T quantum ⁇ T stolen
  • Equation (2) By using Equation (2) to determine VCPU 112 utilization, known demand frequency scaling algorithms for native OS environments may be used with the difference being that VCPU 112 utilization (instead of PCPU 104 utilization) is provided as the input to the DBS-based algorithm 140. Taking VCPU utilization as an input, the CPU-DBS algorithm outputs a lower frequency for low utilization and vice versa.
  • the CPU-DBS algorithm does not provide an optimal balance between performance and power saving. In some instances, better balance may be achieved by profiling additional or different information (other than VCPU 112 utilization) regarding the characteristics of the current workload.
  • additional or different information other than VCPU 112 utilization
  • many hints are generated during virtualization from which workload characteristics may be deduced. More particularly, to consolidate multiple VMs on a single physical machine, the VMM needs to intercept various types of sensitive actions from each VM, emulate the desired behavior, and inject asynchronous virtual interrupts. This new virtualization layer adds overhead to the virtual machine as compared to running on the bare hardware platform. However, the rich hints provided by this inevitable overhead may be used to aid the conjecture of workload characteristics.
  • a workload having input/output (I/O) intensive characteristics may be readily discerned based upon the frequency of I/O requests on the emulated I/O device. If the I/O device is para-virtualized at the driver level, then an enlightened driver could proactively provide statistics that are indicative of the I/O intensive workload.
  • memory virtualization events may be indicators of a memory intensive workload. For instance, the VMM typically may maintain a shadow-page table which is generated from and synchronized to a guest page table once the latter is changed, because the guest page table contains virtual memory addresses which cannot be directly understood by the PCPU. The VMM may mark guest page tables as read-only.
  • a page fault will occur for the VMM to synchronize with. If a VCPU allocates and frees memories frequently, this type of workload may be discerned from statistics indicative of such page faults. Latency sensitive workloads may also be readily discerned. Typically, a latency sensitive pattern is characterized by multiple timers and interrupts. Since the sources of the timers and interrupts are virtualized by the VMM, the VMM can track those events to discern a latency sensitive workload.
  • the virtualization events discussed above are exemplary only and that other types and combinations of events or indicators may be used to profile and deduce characteristics of a workload and, thus, to select an appropriate scaling algorithm.
  • the manner in which workload characteristics may be profiled and deduced may vary depending on the particular implementation of the VMM, the features of the underlying hardware virtualization features, etc. Regardless of the particular manner in which workload characteristics are discerned, embodiments of the invention may import those characteristics into the VCPU-based scaling framework shown in Fig. 1 . Using these workload characteristics, a particular scaling algorithm may be selected for binding to the VCPU instance allocated to the VCPU running that workload.
  • a memory intensive workload typically has many stall cycles. As such, such a workload could run with a lower frequency clock without losing performance, while at the same time, the lower frequency operation could result in power savings.
  • the CPU-DBS algorithm is not a good candidate since the VCPU may be fully saturated under the memory intensive workload and the CPU-DBS algorithm would compute a high frequency.
  • a power savings-based scaling algorithm could be selected based on the recognition that a memory intensive workload can run at a lower frequency.
  • the power savings algorithm may be selected such that a lower frequency operation may result.
  • a latency sensitive workload may have low VCPU utilization.
  • the CPU-DBS algorithm would result in selection of a low frequency.
  • a performance-based scaling algorithm could be selected that would ensure high frequency operation when a latency sensitive workload is deduced from the profiled indicators.
  • a CPU-DBS algorithm may still have many applications. For instance, if few virtualization events are detected, then VCPU utilization may be an appropriate indicator to scale frequency.
  • the profiling channel 138 for each VBFVS instance 136 profiles virtualization events by collecting information from various virtualization event sources, such as a model of a virtual device 146, communications between virtual machines 148, a virtual interrupt controller 150, memory-related virtualization event sources 152, etc.
  • the profiling channel 138 receives information from the scheduler 142 regarding VCPU utilization, including the VCPU's allocated quantum, stolen time, blocked time, etc.
  • the information from these various sources is continuously analyzed during the quantum in which the VCPU 112 is the scheduled-in VCPU so that the VFBVS instance 136 can dynamically bind itself to a particular scaling algorithm 140a-c based on the workload characteristic patterns that are discerned from the information.
  • available scaling algorithms may include a CPU-DBS algorithm 140a, a power-savings algorithm 140b, and a performance-based algorithm 140c, among many others.
  • VCPU-based scaling is not limited to the particular embodiments described herein.
  • VCPU-based scaling may be implemented by simply using VCPU frequency change requests as the profiling information, and either selecting a pass-through scaling algorithm to pass the request through to the VBFVS center or routing the request directly to the VBFVS center.
  • Embodiments may be implemented in software code (including the VMM 110 and any of its components) and may be stored on a storage medium having stored thereon instructions which can be used to program a system to perform the instructions.
  • the storage medium may include, but is not limited to, any type of disk including floppy disks, optical disks, solid state drives, compact disk read-only memories, compact disk rewritables, and magneto-optical disks, semiconductor devices such as read-only memories, random access memories, such as dynamic random access memories, static random access memories, erasable programmable read-only memories, flash memories, electrically erasable programmable read-only memories, magnetic or optical cards, or any other type of media suitable for storing electronic instructions. Instructions of the software may be loaded for execution on a processor (such as the PCPU 104 in Fig. 1 ).

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Software Systems (AREA)
  • Power Sources (AREA)
EP10252026.9A 2009-12-22 2010-11-30 Virtual-CPU based frequency and voltage scaling Not-in-force EP2348410B1 (en)

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
US12/644,359 US8364997B2 (en) 2009-12-22 2009-12-22 Virtual-CPU based frequency and voltage scaling

Publications (3)

Publication Number Publication Date
EP2348410A2 EP2348410A2 (en) 2011-07-27
EP2348410A3 EP2348410A3 (en) 2013-07-10
EP2348410B1 true EP2348410B1 (en) 2018-06-20

Family

ID=43778272

Family Applications (1)

Application Number Title Priority Date Filing Date
EP10252026.9A Not-in-force EP2348410B1 (en) 2009-12-22 2010-11-30 Virtual-CPU based frequency and voltage scaling

Country Status (4)

Country Link
US (1) US8364997B2 (zh)
EP (1) EP2348410B1 (zh)
JP (1) JP5885920B2 (zh)
CN (1) CN102103516B (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
TWI813259B (zh) * 2022-04-19 2023-08-21 凌通科技股份有限公司 在非揮發性隨機存取記憶體上做演算法綁定的方法與使用其之演算法電路

Families Citing this family (29)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2009125789A1 (ja) * 2008-04-09 2009-10-15 日本電気株式会社 計算機システムとその動作方法
JP5435133B2 (ja) * 2010-07-13 2014-03-05 富士通株式会社 情報処理装置、情報処理装置の制御方法及びプログラム
JP5696603B2 (ja) * 2011-06-29 2015-04-08 富士通株式会社 計算機システム、計算機システムの電力制御方法およびプログラム
US8185758B2 (en) * 2011-06-30 2012-05-22 Intel Corporation Method and system for determining an energy-efficient operating point of a platform
WO2012163050A1 (zh) * 2011-11-14 2012-12-06 华为技术有限公司 一种降低芯片功耗的方法及设备
US8892919B2 (en) * 2011-12-14 2014-11-18 Ati Technologies Ulc Method and apparatus for power management of a processor in a virtual environment
US9513688B2 (en) 2013-03-16 2016-12-06 Intel Corporation Measurement of performance scalability in a microprocessor
US9778943B2 (en) * 2013-05-28 2017-10-03 Red Hat Israel, Ltd. Timer based virtual processor scheduling and suspension on physical processor for use of physical processor by other processing
KR102110812B1 (ko) 2013-05-30 2020-05-14 삼성전자 주식회사 멀티 코어 시스템 및 멀티 코어 시스템의 작업 스케줄링 방법
US9195291B2 (en) 2013-06-21 2015-11-24 Apple Inc. Digital power estimator to control processor power consumption
US9304573B2 (en) 2013-06-21 2016-04-05 Apple Inc. Dynamic voltage and frequency management based on active processors
US9195490B2 (en) 2013-07-09 2015-11-24 Empire Technology Development Llc Differential voltage and frequency scaling (DVFS) switch reduction
US9703589B2 (en) 2013-08-26 2017-07-11 Vmware, Inc. Networking stack of virtualization software configured to support latency sensitive virtual machines
CN103488522B (zh) * 2013-09-22 2017-10-17 华为技术有限公司 一种管理虚拟机的设备的方法及设备
US10289437B2 (en) 2014-01-07 2019-05-14 Red Hat Israel, Ltd. Idle processor management in virtualized systems via paravirtualization
US9436263B2 (en) * 2014-02-21 2016-09-06 Qualcomm Incorporated Systems and methods for power optimization using throughput feedback
US10365936B2 (en) 2014-02-27 2019-07-30 Red Hat Israel, Ltd. Idle processor management by guest in virtualized systems
US9606605B2 (en) 2014-03-07 2017-03-28 Apple Inc. Dynamic voltage margin recovery
JP6339240B2 (ja) * 2014-06-24 2018-06-06 インテル コーポレイション 仮想マシン電力管理
KR102169692B1 (ko) * 2014-07-08 2020-10-26 삼성전자주식회사 멀티-코어 프로세서를 포함하는 시스템 온 칩 및 그것의 동적 전력 관리 방법
CN104102327B (zh) * 2014-07-28 2017-02-15 联想(北京)有限公司 一种工作处理器的控制电路和电子设备
WO2016144368A1 (en) * 2015-03-12 2016-09-15 Hewlett Packard Enterprise Development Lp Frequency of a clock signal
US10095597B2 (en) * 2016-01-13 2018-10-09 International Business Machines Corporation Managing a set of wear-leveling data using a set of thread events
US10078457B2 (en) 2016-01-13 2018-09-18 International Business Machines Corporation Managing a set of wear-leveling data using a set of bus traffic
US9886324B2 (en) 2016-01-13 2018-02-06 International Business Machines Corporation Managing asset placement using a set of wear leveling data
EP3338410A4 (en) * 2016-01-18 2018-08-01 Huawei Technologies Co., Ltd. System and method for cloud workload provisioning
US10678603B2 (en) 2016-09-01 2020-06-09 Microsoft Technology Licensing, Llc Resource oversubscription based on utilization patterns in computing systems
US10948957B1 (en) 2019-09-26 2021-03-16 Apple Inc. Adaptive on-chip digital power estimator
CN113687909B (zh) * 2021-07-28 2024-01-30 华东计算技术研究所（中国电子科技集团公司第三十二研究所） 基于微内核的分时vcpu多核调度方法及系统

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20060047987A1 (en) * 2004-08-31 2006-03-02 Rajeev Prabhakaran Dynamic clock frequency adjustment based on processor load
US20090031304A1 (en) * 2007-07-25 2009-01-29 Wei Song Method and device for scheduling true CPU resources for a virtual machine system

Family Cites Families (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US7219241B2 (en) * 2002-11-30 2007-05-15 Intel Corporation Method for managing virtual and actual performance states of logical processors in a multithreaded processor using system management mode
US7739532B2 (en) * 2004-06-07 2010-06-15 Intel Corporation Method, apparatus and system for enhanced CPU frequency governers
US7386739B2 (en) * 2005-05-03 2008-06-10 International Business Machines Corporation Scheduling processor voltages and frequencies based on performance prediction and power constraints
US7861068B2 (en) * 2006-03-07 2010-12-28 Intel Corporation Method and apparatus for using dynamic workload characteristics to control CPU frequency and voltage scaling
US7730340B2 (en) * 2007-02-16 2010-06-01 Intel Corporation Method and apparatus for dynamic voltage and frequency scaling
CN101470634B (zh) * 2007-12-24 2012-05-09 联想(北京)有限公司 Cpu资源调度方法及cpu资源调度装置
US8370663B2 (en) * 2008-02-11 2013-02-05 Nvidia Corporation Power management with dynamic frequency adjustments
US8892916B2 (en) * 2008-08-06 2014-11-18 International Business Machines Corporation Dynamic core pool management
US20110153984A1 (en) * 2009-12-21 2011-06-23 Andrew Wolfe Dynamic voltage change for multi-core processing

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20060047987A1 (en) * 2004-08-31 2006-03-02 Rajeev Prabhakaran Dynamic clock frequency adjustment based on processor load
US20090031304A1 (en) * 2007-07-25 2009-01-29 Wei Song Method and device for scheduling true CPU resources for a virtual machine system

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
INTEL OPEN SOURCE TECHNOLOGY CENTER: "How virtualization makes power management different", PROCEEDINGS OF THE LINUX SYMPOSIUM, 27 June 2007 (2007-06-27), Ottawa, Ontario, pages 205 - 214, XP055208120 *

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
TWI813259B (zh) * 2022-04-19 2023-08-21 凌通科技股份有限公司 在非揮發性隨機存取記憶體上做演算法綁定的方法與使用其之演算法電路

Also Published As

Publication number Publication date
CN102103516A (zh) 2011-06-22
EP2348410A2 (en) 2011-07-27
US20110154321A1 (en) 2011-06-23
US8364997B2 (en) 2013-01-29
JP5885920B2 (ja) 2016-03-16
JP2011134333A (ja) 2011-07-07
EP2348410A3 (en) 2013-07-10
CN102103516B (zh) 2015-04-08

Similar Documents

Publication Publication Date Title
EP2348410B1 (en) Virtual-CPU based frequency and voltage scaling
Shen et al. Cloudscale: elastic resource scaling for multi-tenant cloud systems
Dhiman et al. vGreen: a system for energy efficient computing in virtualized environments
EP2071458B1 (en) Power control method for virtual machine and virtual computer system
US8489904B2 (en) Allocating computing system power levels responsive to service level agreements
Dhiman et al. vgreen: A system for energy-efficient management of virtual machines
US20080301473A1 (en) Method and system for hypervisor based power management
US20140351613A1 (en) Virtual machine power consumption measurement and management
US20100191854A1 (en) Process demand prediction for distributed power and resource management
Kwon et al. Virtualizing performance asymmetric multi-core systems
Verma et al. Cosmig: Modeling the impact of reconfiguration in a cloud
US20140373010A1 (en) Intelligent resource management for virtual machines
US20080028411A1 (en) System and method for controlling aggregate CPU usage by virtual machines and driver domains over a plurality of scheduling intervals
WO2013097068A1 (en) Scheduling virtual central processing units of virtual machines among physical processing units
Broyles et al. IBM energyscale for POWER7 processor-based systems
US8032882B2 (en) System and method for controlling aggregate CPU usage by virtual machines and driver domains
Gifford et al. Dna: Dynamic resource allocation for soft real-time multicore systems
Kleineweber et al. Rule-based mapping of virtual machines in clouds
Xilong et al. An energy-efficient virtual machine scheduler based on CPU share-reclaiming policy
Wang et al. Energy-efficient virtual machine scheduling in performance-asymmetric multi-core architectures
US10474392B2 (en) Dynamic scheduling for virtual storage devices
Kamga et al. Extended scheduler for efficient frequency scaling in virtualized systems
Kang et al. VIP: Virtual performance-state for efficient power management of virtual machines
Simão et al. A classification of middleware to support virtual machines adaptability in IaaS
Zhang CoolCloud: Improving energy efficiency in virtualized data centers

Legal Events

Date Code Title Description
PUAI Public reference made under article 153(3) epc to a published international application that has entered the european phase

Free format text: ORIGINAL CODE: 0009012

17P Request for examination filed

Effective date: 20101220

AK Designated contracting states

Kind code of ref document: A2

Designated state(s): AL AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HR HU IE IS IT LI LT LU LV MC MK MT NL NO PL PT RO RS SE SI SK SM TR

AX Request for extension of the european patent

Extension state: BA ME

PUAL Search report despatched

Free format text: ORIGINAL CODE: 0009013

AK Designated contracting states

Kind code of ref document: A3

Designated state(s): AL AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HR HU IE IS IT LI LT LU LV MC MK MT NL NO PL PT RO RS SE SI SK SM TR

AX Request for extension of the european patent

Extension state: BA ME

RIC1 Information provided on ipc code assigned before grant

Ipc: G06F 9/50 20060101AFI20130603BHEP

Ipc: G06F 1/32 20060101ALI20130603BHEP

17Q First examination report despatched

Effective date: 20130628

GRAP Despatch of communication of intention to grant a patent

Free format text: ORIGINAL CODE: EPIDOSNIGR1

STAA Information on the status of an ep patent application or granted ep patent

Free format text: STATUS: GRANT OF PATENT IS INTENDED

INTG Intention to grant announced

Effective date: 20180124

GRAS Grant fee paid

Free format text: ORIGINAL CODE: EPIDOSNIGR3

GRAA (expected) grant

Free format text: ORIGINAL CODE: 0009210

STAA Information on the status of an ep patent application or granted ep patent

Free format text: STATUS: THE PATENT HAS BEEN GRANTED

AK Designated contracting states

Kind code of ref document: B1

Designated state(s): AL AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HR HU IE IS IT LI LT LU LV MC MK MT NL NO PL PT RO RS SE SI SK SM TR

REG Reference to a national code

Ref country code: GB

Ref legal event code: FG4D

REG Reference to a national code

Ref country code: IE

Ref legal event code: FG4D

REG Reference to a national code

Ref country code: AT

Ref legal event code: REF

Ref document number: 1011083

Country of ref document: AT

Kind code of ref document: T

Effective date: 20180715

REG Reference to a national code

Ref country code: DE

Ref legal event code: R096

Ref document number: 602010051383

Country of ref document: DE

REG Reference to a national code

Ref country code: NL

Ref legal event code: FP

PG25 Lapsed in a contracting state [announced via postgrant information from national office to epo]

Ref country code: FI

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20180620

Ref country code: NO

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20180920

Ref country code: LT

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20180620

Ref country code: SE

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20180620

Ref country code: BG

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20180920

REG Reference to a national code

Ref country code: LT

Ref legal event code: MG4D

PG25 Lapsed in a contracting state [announced via postgrant information from national office to epo]

Ref country code: LV

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20180620

Ref country code: RS

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20180620

Ref country code: GR

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20180921

Ref country code: HR

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20180620

REG Reference to a national code

Ref country code: AT

Ref legal event code: MK05

Ref document number: 1011083

Country of ref document: AT

Kind code of ref document: T

Effective date: 20180620

PG25 Lapsed in a contracting state [announced via postgrant information from national office to epo]

Ref country code: IS

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20181020

Ref country code: AT

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20180620

Ref country code: EE

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20180620

Ref country code: PL

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20180620

Ref country code: SK

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20180620

Ref country code: CZ

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20180620

Ref country code: RO

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20180620

PG25 Lapsed in a contracting state [announced via postgrant information from national office to epo]

Ref country code: ES

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20180620

Ref country code: IT

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20180620

Ref country code: SM

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20180620

REG Reference to a national code

Ref country code: DE

Ref legal event code: R097

Ref document number: 602010051383

Country of ref document: DE

PLBE No opposition filed within time limit

Free format text: ORIGINAL CODE: 0009261

STAA Information on the status of an ep patent application or granted ep patent

Free format text: STATUS: NO OPPOSITION FILED WITHIN TIME LIMIT

26N No opposition filed

Effective date: 20190321

PG25 Lapsed in a contracting state [announced via postgrant information from national office to epo]

Ref country code: DK

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20180620

REG Reference to a national code

Ref country code: CH

Ref legal event code: PL

PG25 Lapsed in a contracting state [announced via postgrant information from national office to epo]

Ref country code: MC

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20180620

Ref country code: LU

Free format text: LAPSE BECAUSE OF NON-PAYMENT OF DUE FEES

Effective date: 20181130

REG Reference to a national code

Ref country code: BE

Ref legal event code: MM

Effective date: 20181130

REG Reference to a national code

Ref country code: IE

Ref legal event code: MM4A

PG25 Lapsed in a contracting state [announced via postgrant information from national office to epo]

Ref country code: LI

Free format text: LAPSE BECAUSE OF NON-PAYMENT OF DUE FEES

Effective date: 20181130

Ref country code: SI

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20180620

Ref country code: CH

Free format text: LAPSE BECAUSE OF NON-PAYMENT OF DUE FEES

Effective date: 20181130

PG25 Lapsed in a contracting state [announced via postgrant information from national office to epo]

Ref country code: FR

Free format text: LAPSE BECAUSE OF NON-PAYMENT OF DUE FEES

Effective date: 20181130

Ref country code: IE

Free format text: LAPSE BECAUSE OF NON-PAYMENT OF DUE FEES

Effective date: 20181130

PG25 Lapsed in a contracting state [announced via postgrant information from national office to epo]

Ref country code: BE

Free format text: LAPSE BECAUSE OF NON-PAYMENT OF DUE FEES

Effective date: 20181130

Ref country code: AL

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20180620

PG25 Lapsed in a contracting state [announced via postgrant information from national office to epo]

Ref country code: MT

Free format text: LAPSE BECAUSE OF NON-PAYMENT OF DUE FEES

Effective date: 20181130

PG25 Lapsed in a contracting state [announced via postgrant information from national office to epo]

Ref country code: TR

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20180620

PG25 Lapsed in a contracting state [announced via postgrant information from national office to epo]

Ref country code: PT

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20180620

PG25 Lapsed in a contracting state [announced via postgrant information from national office to epo]

Ref country code: HU

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT; INVALID AB INITIO

Effective date: 20101130

Ref country code: CY

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20180620

Ref country code: MK

Free format text: LAPSE BECAUSE OF NON-PAYMENT OF DUE FEES

Effective date: 20180620

PGFP Annual fee paid to national office [announced via postgrant information from national office to epo]

Ref country code: GB

Payment date: 20200925

Year of fee payment: 11

Ref country code: NL

Payment date: 20200924

Year of fee payment: 11

PGFP Annual fee paid to national office [announced via postgrant information from national office to epo]

Ref country code: DE

Payment date: 20200922

Year of fee payment: 11

REG Reference to a national code

Ref country code: DE

Ref legal event code: R119

Ref document number: 602010051383

Country of ref document: DE

REG Reference to a national code

Ref country code: NL

Ref legal event code: MM

Effective date: 20211201

GBPC Gb: european patent ceased through non-payment of renewal fee

Effective date: 20211130

PG25 Lapsed in a contracting state [announced via postgrant information from national office to epo]

Ref country code: NL

Free format text: LAPSE BECAUSE OF NON-PAYMENT OF DUE FEES

Effective date: 20211201

PG25 Lapsed in a contracting state [announced via postgrant information from national office to epo]

Ref country code: GB

Free format text: LAPSE BECAUSE OF NON-PAYMENT OF DUE FEES

Effective date: 20211130

Ref country code: DE

Free format text: LAPSE BECAUSE OF NON-PAYMENT OF DUE FEES

Effective date: 20220601